非线性互补问题的内点法  

Interior point methods for NCP

作　　者：何郁波[1] 梁茜[1] 田亚娟[1] 马昌凤[1] 

机构地区：[1]桂林电子科技大学计算科学与数学系,广西桂林541004

出　　处：《桂林电子工业学院学报》2006年第3期207-211,共5页Journal of Guilin Institute of Electronic Technology

基　　金：广西自然科学基金(0640165)

摘　　要：在利用惩罚函数法求解非线性互补问题的基础上,将非线性互补问题转化成不等式约束问题,利用引入的附加变量,将其转化成等式约束问题,再采用将罚函数法和L agrange乘子法相结合的方法进行求解,算法的收敛性得到了证明,数值实验与仅用惩罚函数方法相比,有更强的收敛性和更快的收敛速度。In this paper we first discuss the basis for the transformation of nonlinear complementarity problems into the nonlinear programming problems with inequality constraints. By using additive variables, we convert the inequality constraints problems into a sequence of nonlinear programming problems with equality constraints. We then propose the penalty function methods and Lagrange multiply methods for solving the nonlinear programming problems with equality constraints. The convergence of the algorithm is also proved. Numerical example is given to illustrate the method.

关 键 词：互补问题 罚函数法 LAGRANGE乘子法 收敛性 

分 类 号：O224.2[理学—运筹学与控制论]